THE MODELING OF SPILLS RESULTING FROM THE CATASTROPHIC FAILURE OF ABOVE GROUND STORAGE TANKS AND THE DEVELOPMENT OF MITIGATION

Abstract

ABSTRACT The risk of accidents involving the catastrophic failure of storage tanks is estimated to be low, in the region of 5 × l0−6 per tank year. However, recent accidents involving major oil spills at storage facilities located in Belgium (2004) along with USA and England (2005) have shown that tank failures do nevertheless occur. Causalities of such events vary; the consequences however are ordinarily the same, incurring environmental, financial and infrastructure losses. The normal mitigation technique employed to prevent such losses is secondary containment, usually in the form of a bund wall or earthen dyke. Researchers have investigated the reliability of such methods, examining the effects of tank failure, both theoretically and experimentally in terms of loss of containment. A United Kingdom Health and Safety Executive (HSE) review conducted in 1997 concluded that the then available data was limited and focussed attention on the work of Greenspan and Johansson (1981) and the later work of Trobojevic and Slater (1989). This led to the HSE commissioning Liverpool John Moores University (LIMU) in 2003 to undertake a large-scale spill-modelling program with the aim of quantifying the level of overtopping and the magnitudes of the dynamic pressures on the bunds. The study examined the effect of axisymmetric releases on a total of 96 tank and bund arrangements. Such losses have proven to be significant and in some cases the nature of the dynamic pressures has brought in to question the structural integrity of the bunds themselves. Research has since concentrated on modelling alternative modes of failure, such as directional releases, which could be considered to be the more common mode of failure likely to be encountered. The conclusions to this work have generated additional research to investigate possible methods of mitigation that could be incorporated into the design of facilities with the ultimate aim of further reducing losses in the event of tank failure. Two promising methods have been identified, one involving modification to the primary containment (tank) with the other being a change to the design of the profile of the secondary containment (bund wall).